CN103886564A - PET heart image myocardial contour line segmentation method and device - Google Patents

Abstract

The embodiment of the invention discloses a PET heart image myocardial contour line segmentation method comprising that: coarse segmentation is performed on the received PET heart image so that initial contour lines of a myocardial area are obtained; a derivative order alpha in curve length regular terms in a function description of a level set image segmentation model is arranged to be a fractional order; curve evolution calculation is performed on the initial contour lines by using the level set image segmentation model and selecting at least three pixel points adjacent to each pixel point on the initial contour lines; and the obtained curve evolution calculation results act as the contour lines of the PET heart image myocardial area. Therefore, the derivative order in the curve length regular terms is arranged to be the fractional order so that any number of multiple adjacent pixel points can be selected to act as references of the curve evolution calculation by aiming at each point on the initial contour lines when the curve evolution calculation is performed, and thus boundary accuracy of the myocardial area finally obtained via calculation can be substantially enhanced.

Description

A kind of PET cardiac image Myocardial Contour segmentation method and apparatus

Technical field

The present invention relates to medical instruments field, particularly relate to a kind of PET cardiac image Myocardial Contour segmentation method and apparatus.

Background technology

At present, positron emission computer fault imaging (Positron Emission Computed Tomography, PET) system or PET/CT system demonstrate remarkable performance in tumour, cardiovascular and nervous system area research, and its myocardial metabolic imaging can be used for effectively determining and differentiating cardiac muscle cell's activity.

Carrying out in the process of myocardial metabolic imaging, need to be in shown PET cardiac image depicting the boundary accurate of myocardial region, this is prerequisite and the committed step of carrying out subsequent treatment according to myocardial metabolic imaging, and the whether accurate accuracy that can greatly have influence on the result of subsequent treatment of its rendering results.

The method of at present PET cardiac image being carried out to myocardial region contour description is mainly divided into two steps, first step is first means by manual observation, from PET cardiac image, myocardial region is roughly determined, be that is to say the initial profile that obtains myocardial region by coarse segmentation.Second step is now main, and what use is that the outline line of the initial profile that first step is obtained of the image partition method based on level set (Level set) carries out curve EVOLUTIONARY COMPUTATION, obtain thus the border of accurate myocardial region, this image partition method need to use level set Image Segmentation Model and carry out Functional Analysis, in the function representation of level set Image Segmentation Model, all comprise two parts of length of curve regular terms and data fitting item, wherein length of curve regular terms is used for carrying out curve EVOLUTIONARY COMPUTATION, in the time carrying out curve EVOLUTIONARY COMPUTATION, one or two pixels that can and can only rely on the fixed number that on initial profile line, each pixel is adjacent in prior art calculate, make the boundary accurate of the final myocardial region obtaining lower.

Summary of the invention

In order to solve the problems of the technologies described above, the invention provides a kind of PET cardiac image Myocardial Contour segmentation method and apparatus, improve to calculate in evolution curve process and can only obtain the mechanism of fixed qty neighbor pixel as computing reference, improved thus the precision of final calculating.

The embodiment of the invention discloses following technical scheme:

A kind of PET cardiac image Myocardial Contour segmentation method, comprising:

The PET cardiac image receiving is carried out to coarse segmentation, obtain the initial profile line of myocardial region;

Derivative exponent number α in length of curve regular terms in the function representation of level set Image Segmentation Model is set to fractional order;

Use described level set Image Segmentation Model and choose at least three pixels that on described initial profile line, each pixel is adjacent described initial profile line is carried out to curve EVOLUTIONARY COMPUTATION;

Outline line using the curve evolvement result of calculation obtaining as described PET cardiac image Myocardial region.

Preferably,

Described level set Image Segmentation Model comprises unskirted active profile C-V(Chan-Vese) model or based on region extendible matching RSF(Region-Scalable Fitting) model.

Preferably,

Described derivative exponent number α is greater than 0 to be less than 2 and be not equal to 1 rational number, irrational number or plural number.

Preferably, described the PET cardiac image receiving is carried out to coarse segmentation, is specially:

Use iteration threshold method to carry out coarse segmentation to the PET cardiac image receiving.

Preferably, at least three pixels that on the described level set Image Segmentation Model of described use and described initial profile line, each pixel is adjacent carry out curve EVOLUTIONARY COMPUTATION to described initial profile line, are specially:

When the iteration result of calculating when described curve evolvement reaches preset times for convergence or iterations, the curve evolvement result of calculation obtaining.

A kind of PET cardiac image Myocardial Contour segmentation device, comprising:

Initial profile line cutting unit, carries out coarse segmentation for the PET cardiac image to receiving, and obtains the initial profile line of myocardial region;

Fractional order setting unit, is set to fractional order for the derivative exponent number α in the length of curve regular terms of the function representation of level set Image Segmentation Model;

Curve evolvement computing unit, carries out curve EVOLUTIONARY COMPUTATION for using described level set Image Segmentation Model and choosing at least three pixels that on described initial profile line, each pixel is adjacent to described initial profile line;

Outline line acquiring unit, for the outline line using the curve evolvement result of calculation obtaining as described PET cardiac image Myocardial region.

Preferably,

Described level set Image Segmentation Model comprises unskirted active profile C-V(Chan-Vese) model or based on region extendible matching RSF(Region-Scalable Fitting) model.

Preferably,

Described derivative exponent number α is greater than 0 to be less than 2 and be not equal to 1 rational number, irrational number or plural number.

Preferably, described initial profile line cutting unit, is specially:

Use iteration threshold method to carry out coarse segmentation to the PET cardiac image receiving.

Preferably, described curve evolvement computing unit is specially:

When the iteration result of calculating when described curve evolvement reaches preset times for convergence or iterations, the curve evolvement result of calculation obtaining.

Can be found out by technique scheme, derivative exponent number by the derivative to level set function in the length of curve regular terms in the function representation of level set Image Segmentation Model is set to fractional order, make thus when the initial profile line that uses the myocardial region of this level set Image Segmentation Model to coarse segmentation carries out curve EVOLUTIONARY COMPUTATION, for the point on each initial profile line, all can choose the reference that any number of neighbor pixels calculate as curve evolvement, and be elected to when being taken to few three neighbor pixels, can improve significantly the boundary accurate of the myocardial region of final calculating acquisition.

Brief description of the drawings

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, to the accompanying drawing of required use in embodiment or description of the Prior Art be briefly described below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, do not paying under the prerequisite of creative work, can also obtain according to these accompanying drawings other accompanying drawing.

Fig. 1 is the method flow diagram of a kind of PET cardiac image of the present invention Myocardial Contour segmentation method;

The myocardial region outline line schematic diagram of Fig. 2 a for obtaining according to C-V model curve EVOLUTIONARY COMPUTATION;

The myocardial region outline line schematic diagram that Fig. 2 b obtains according to fractional order C-V model curve EVOLUTIONARY COMPUTATION for the present invention;

The myocardial region outline line schematic diagram of Fig. 2 c for obtaining according to RSF model curve EVOLUTIONARY COMPUTATION;

The myocardial region outline line schematic diagram that Fig. 2 d obtains according to fractional order RSF model curve EVOLUTIONARY COMPUTATION for the present invention;

Fig. 3 is the structure drawing of device of a kind of PET cardiac image of the present invention Myocardial Contour segmentation device.

Embodiment

The embodiment of the present invention provides a kind of PET cardiac image Myocardial Contour segmentation method and apparatus.On the one hand, derivative exponent number by the derivative to level set function in the length of curve regular terms in the function representation of level set Image Segmentation Model is set to fractional order, make thus when the initial profile line that uses the myocardial region of this level set Image Segmentation Model to coarse segmentation carries out curve EVOLUTIONARY COMPUTATION, for the point on each initial profile line, all can choose the reference that any number of neighbor pixels calculate as curve evolvement, and be elected to when being taken to few three neighbor pixels, can improve significantly the boundary accurate of the myocardial region of final calculating acquisition.

On the other hand, use iteration threshold method to replace artificial means originally to dock the PET cardiac image of receiving and carry out coarse segmentation, further improved the Efficiency and accuracy that obtains myocardial region initial profile line.

For above-mentioned purpose of the present invention, feature and advantage can be become apparent more, below in conjunction with accompanying drawing, the embodiment of the present invention is described in detail.

Embodiment mono-

In technical scheme of the present invention, image partition method fractional calculus is theoretical and level set combines, propose the concept that fractional order level set image is cut apart, and constructed two Image Segmentation Model based on fractional order level set, that is: fractional order C-V model and fractional order RSF model.And utilize these two models to cut apart PET cardiac image, obtain than traditional better segmentation result of the Image Segmentation Model based on level set.

Refer to Fig. 1, its method flow diagram that is a kind of PET cardiac image of the present invention Myocardial Contour segmentation method, the method comprises the following steps:

S101: the PET cardiac image receiving is carried out to coarse segmentation, obtain the initial profile line of myocardial region;

In prior art, receiving after a PET cardiac image, is generally that first Preliminary division goes out the general profile of myocardial region by after artificial observation, and this coarse segmentation efficiency is low, and dividing precision and operator's working experience has direct relation.And in embodiments of the present invention, except carrying out follow-up curve evolvement calculating for mark off initial profile line in the situation that this people, a kind of method of preferably obtaining myocardial region initial profile line has also been proposed, by the gray scale in PET cardiac image is analyzed, use iteration threshold method to carry out coarse segmentation to the PET cardiac image receiving, described iteration threshold method is specially:

Step 1: initial threshold T(Theshhold is set), wherein T can be made as the maximum gradation value in image pixel and the mean value of minimum gradation value in this PET cardiac image; Iteration is set and upgrades threshold value TT, iteration termination rules d=|T-TT|<0.5 is set.

Step 2: the PET cardiac image receiving is divided into two region R1 and R2 according to described initial threshold T, the rule of dividing is generally, in R1 region can be the set that gray-scale value is more than or equal to all pixels of predetermined threshold value T, in R2 region, can be the set that gray-scale value is less than all pixels of predetermined threshold value T;

Step 3: the average gray value T2 of all pixels in average gray value T1 and the R2 of all pixels in calculating R1;

Step 4: iteration is set and upgrades threshold value TT=0.5* (T1+T2), calculate the value of d=|T-TT|;

Step 5: if the value of d is more than or equal to 0.5, using the value of current TT as initial threshold T, and repeat step 2 to step 4, successive iteration is until meet iteration termination rules d=|T-TT|<0.5

The efficiency that the myocardial region initial profile line that iteration threshold method obtains is thus compared artificial division is higher, and accuracy also can be more stable.

S102: the derivative exponent number α in the length of curve regular terms in the function representation of level set Image Segmentation Model is set to fractional order;

Here need the derivative exponent number α to being arranged to fractional order to describe, in general, scope to the derivative exponent number arranging does not limit, but under the prerequisite of conveniently calculating, derivative exponent number α can be arranged between 0 to 2, certainly can not be integer 1, can be rational number, irrational number or plural number within the scope of this.In the time that derivative exponent number α is arranged in this scope, can make only need to carry out just can obtain after relatively less iterations the result of calculation of final pinpoint accuracy.

Level set Image Segmentation Model described in the embodiment of the present invention mainly refers to the level set Image Segmentation Model that comprises integer order derivative in its length of curve regular terms, and now the most frequently used level set Image Segmentation Model mainly comprises C-V(Chan-Vese) model or can scaling region matching RSF(Region-Scalable Fitting) model.

For C-V model, its function representation is:

E ^cV(c ₁, c ₂, C) and=μ Length (C)+data fitting item,

For RSF model, its function representation is:

E ^rSF(f ₁(X), f ₂(X), C)=μ Length (C)+data fitting item,

Wherein Length (C) is length of curve regular terms, and the derivative that comprises integer rank, when it is applied green theorem and is introduced after level set function φ, obtains its equation expression for level set:

$\mathrm{Length}\{\mathrm{\&phi;}=0\}={\&Integral;}_{\mathrm{\&Omega;}}{\mathrm{\&delta;}}_3\left(\mathrm{\&phi;}(x,y)\right)|\&dtri;\mathrm{\&phi;}(x,y)|\mathrm{dxdy}$

Wherein:

for the first order derivative of level set function φ (x, y).

When its derivative exponent number, α is set to after fractional order, and length of curve regular terms specifically becomes:

Length ^α{φ=0}=∫ _Ωδ _ε(φ(x,y))|D ^αφ(x,y)|dxdy

Wherein:

D ^αφ (x, y) is the α order derivative of level set function φ (x, y).

That is to say, the curvilinear integral under first order derivative has originally been become to the curve surface integral in the embodiment of the present invention.Of the prior art

belong to integer ladder degree.Carry out during image cuts apart utilizing Level Set Method, level set function φ (x, that integer ladder degree y) represents is each pixel (x, y) level set function φ (x, y) only depend on the level set function φ (x of its adjacent pixel or two pixels, y), that is to say the First-order Gradient of level set function φ (x, y) only there is local property, in other words, in the case of can for the pixel number of reference fix and little, final result of calculation is certain to cause certain limitation.And if by the First-order Gradient of level set function

expand to fractional order gradient (derivative) D ^αφ (x, y), because fractional order differential has the characteristic of " long memory (long-memory) ", the level set function φ (x of a pixel, y) fractional order differential not merely depends on the level set function φ (x that puts one or two adjacent pixel with this, y), but with the level set function φ (x of its all pixels around, y) relevant, here all pixels of indication can be the set of containing all pixels of entire image, that is to say fractional order gradient (derivative) D of level set function φ (x, y) ^αφ (x, y) has global nature, and in other words, in the case of can contain whole the pixel in PET cardiac image for the pixel number of reference, it is more accurate that final result of calculation is certain to.If but calculated with reference to all pixels, can make calculated amount excessive, even if finally can obtain the result of calculation of pinpoint accuracy, but the time of waiting for result of calculation can be also very long, therefore in the case of final computational accuracy not being produced considerable influence, in numerical evaluation, in order to reduce calculated amount, usually the technical limitation of choosing the pixel for calculating is arrived in certain scope, in embodiments of the present invention, being set to is at least at least three adjacent pixels of each pixel on initial profile line.

S103: use described level set Image Segmentation Model and choose at least three pixels that on described initial profile line, each pixel is adjacent described initial profile line is carried out to curve EVOLUTIONARY COMPUTATION;

The computation process of calculating for curve evolvement, integer order derivative in curve evolvement regular terms is changed to Fractional Derivative, other computation process is identical with prior art, in the time that calculating has the function representation of level set Image Segmentation Model of Fractional Derivative, when generally the iterations in the time that convergence appears in result of calculation or in calculating reaches preset times, judge that result of calculation is at this moment final curve evolvement result of calculation.Generally can this preset times be set to iterations 30 times, can certainly be set to other number of times, the present invention does not limit this.

S104: the outline line using the curve evolvement result of calculation obtaining as described PET cardiac image Myocardial region.

Next the fractional order image partition method that uses existing level set image segmentation method and the present invention to propose same width PET cardiac image being carried out to image cuts apart, to obtain the myocardial region outline line in image, the difference of clearer displaying prior art and technical solution of the present invention thus.Wherein, the initial profile line that is used for this PET cardiac image myocardial region of calculating is all also identical, can, by people for marking off, also can calculate by iteration threshold method of the present invention.

After the comparison of cutting apart for the image of same width PET cardiac image, for the fractional order C-V model in level set Image Segmentation Model C-V model and the present invention, the myocardial region outline line obtaining is respectively as shown in Fig. 2 a, 2b, the wherein myocardial region outline line schematic diagram of Fig. 2 a for obtaining according to C-V model curve EVOLUTIONARY COMPUTATION, in figure, the outer contour of dark part is exactly the myocardial region outline line obtaining by C-V model curve EVOLUTIONARY COMPUTATION.The myocardial region outline line schematic diagram that described Fig. 2 b obtains according to fractional order C-V model curve EVOLUTIONARY COMPUTATION for the present invention, in figure, the outer contour of dark part is exactly the myocardial region outline line obtaining by fractional order C-V model curve EVOLUTIONARY COMPUTATION.Comparison can be found out, obviously more accurate with the myocardial region outline line that technical scheme of the present invention was obtained.

After carrying out cutting apart for three times for three width PET cardiac images, obtain the result of calculation of following table:

Can find out, in the time using technical scheme of the present invention, fractional order C-V model obtains from start to finish myocardial region outline line iterations used and is obviously less than in prior art and uses C-V model to cut apart the iterations of cardiac image, that is to say, use after technical scheme of the present invention, not only computational accuracy has obtained significantly improving, and does not also increase in calculated amount, significantly reduce on the contrary, further improved counting yield.That is to say, in the time using fractional order image partition method of the present invention, in the time that fractional order is set to proper, an appropriate fractional-order, just can be in obtaining good segmentation result, effectively reduce the iterations of curve evolvement in calculating.

After the comparison of cutting apart for the image of same width PET cardiac image, for the fractional order RSF model in level set Image Segmentation Model RSF model and the present invention, the myocardial region outline line obtaining is respectively as shown in Fig. 2 c, 2d, the wherein myocardial region outline line schematic diagram of Fig. 2 c for obtaining according to RSF model curve EVOLUTIONARY COMPUTATION, in figure, the outer contour of dark part is exactly the myocardial region outline line obtaining by RSF model curve EVOLUTIONARY COMPUTATION.The myocardial region outline line schematic diagram that Fig. 2 d obtains according to fractional order RSF model curve EVOLUTIONARY COMPUTATION for the present invention, in figure, the outer contour of dark part is exactly the myocardial region outline line obtaining by fractional order RSF model curve EVOLUTIONARY COMPUTATION.Comparison can be found out, has even occurred two-part myocardial region outline line in Fig. 2 c, but obviously lower left corner part belong to error calculate cause, so the myocardial region outline line that technical scheme of the present invention obtains is obviously more accurate.

Can be found out by the present embodiment, derivative exponent number by the derivative to level set function in the length of curve regular terms in the function representation of level set Image Segmentation Model is set to fractional order, make thus when the initial profile line that uses the myocardial region of this level set Image Segmentation Model to coarse segmentation carries out curve EVOLUTIONARY COMPUTATION, for the point on each initial profile line, all can choose the reference that any number of neighbor pixels calculate as curve evolvement, and be elected to when being taken to few three neighbor pixels, can improve significantly the boundary accurate of the myocardial region of final calculating acquisition.

On the other hand, use iteration threshold method to replace artificial means originally to dock the PET cardiac image of receiving and carry out coarse segmentation, further improved the Efficiency and accuracy that obtains myocardial region initial profile line.

Embodiment bis-

Corresponding with a kind of above-mentioned PET cardiac image Myocardial Contour segmentation method, the embodiment of the present invention also provides a kind of PET cardiac image Myocardial Contour segmentation device.Refer to Fig. 3, its structure drawing of device that is a kind of PET cardiac image of the present invention Myocardial Contour segmentation device, this device comprises initial profile line cutting unit 301, fractional order setting unit 302, curve evolvement computing unit 303 and outline line acquiring unit 304:

Initial profile line cutting unit 301, carries out coarse segmentation for the PET cardiac image to receiving, and obtains the initial profile line of myocardial region;

Preferably, described initial profile line cutting unit, is specially:

Use iteration threshold method to carry out coarse segmentation to the PET cardiac image receiving.

Fractional order setting unit 302, is set to fractional order for the derivative exponent number α in the length of curve regular terms of the function representation of level set Image Segmentation Model;

Preferably, described level set Image Segmentation Model comprises unskirted active profile C-V(Chan-Vese) model or based on region extendible matching RSF(Region-Scalable Fitting) model.

Preferably, described derivative exponent number α is greater than 0 to be less than 2 and be not equal to 1 rational number, irrational number or plural number.

Preferably, the length of curve regular terms in the function representation of described level set Image Segmentation Model is specially:

Length ^α{φ=0}=∫ _Ωδ _ε(φ(x,y))|D ^αφ(x,y)|dxdy

Wherein:

D ^αφ (x, y) is the α order derivative of level set function φ (x, y).

Curve evolvement computing unit 303, carries out curve EVOLUTIONARY COMPUTATION for using described level set Image Segmentation Model and choosing at least three pixels that on described initial profile line, each pixel is adjacent to described initial profile line;

Preferably, described curve evolvement computing unit is specially:

When the iteration result of calculating when described curve evolvement reaches preset times for convergence or iterations, the curve evolvement result of calculation obtaining.

Outline line acquiring unit 304, for the outline line using the curve evolvement result of calculation obtaining as described PET cardiac image Myocardial region.

As can be seen from the above-described embodiment, derivative exponent number by the derivative to level set function in the length of curve regular terms in the function representation of level set Image Segmentation Model is set to fractional order, make thus when the initial profile line that uses the myocardial region of this level set Image Segmentation Model to coarse segmentation carries out curve EVOLUTIONARY COMPUTATION, for the point on each initial profile line, all can choose the reference that any number of neighbor pixels calculate as curve evolvement, and be elected to when being taken to few three neighbor pixels, can improve significantly the boundary accurate of the myocardial region of final calculating acquisition.

On the other hand, use iteration threshold method to replace artificial means originally to dock the PET cardiac image of receiving and carry out coarse segmentation, further improved the Efficiency and accuracy that obtains myocardial region initial profile line.

It should be noted that, one of ordinary skill in the art will appreciate that all or part of flow process realizing in above-described embodiment method, can carry out the hardware that instruction is relevant by computer program to complete, described program can be stored in a computer read/write memory medium, this program, in the time carrying out, can comprise as the flow process of the embodiment of above-mentioned each side method.Wherein, described storage medium can be magnetic disc, CD, read-only store-memory body (Read-Only Memory, ROM) or random store-memory body (Random Access Memory, RAM) etc.

Above a kind of PET cardiac image Myocardial Contour segmentation method and apparatus provided by the present invention is described in detail, applied specific embodiment herein principle of the present invention and embodiment are set forth, the explanation of above embodiment is just for helping to understand method of the present invention and core concept thereof; , for one of ordinary skill in the art, according to thought of the present invention, all will change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention meanwhile.

Claims (10)

1. a PET cardiac image Myocardial Contour segmentation method, is characterized in that, comprising:

The PET cardiac image receiving is carried out to coarse segmentation, obtain the initial profile line of myocardial region;

Derivative exponent number α in length of curve regular terms in the function representation of level set Image Segmentation Model is set to fractional order;

Use described level set Image Segmentation Model and choose at least three pixels that on described initial profile line, each pixel is adjacent described initial profile line is carried out to curve EVOLUTIONARY COMPUTATION;

Outline line using the curve evolvement result of calculation obtaining as described PET cardiac image Myocardial region.

2. method according to claim 1, is characterized in that,

Described level set Image Segmentation Model comprises unskirted active profile C-V(Chan-Vese) model or based on region extendible matching RSF(Region-Scalable Fitting) model.

3. method according to claim 1, is characterized in that,

Described derivative exponent number α is greater than 0 to be less than 2 and be not equal to 1 rational number, irrational number or plural number.

4. method according to claim 1, is characterized in that, described the PET cardiac image receiving is carried out to coarse segmentation, is specially:

Use iteration threshold method to carry out coarse segmentation to the PET cardiac image receiving.

5. method according to claim 1, is characterized in that, at least three pixels that on the described level set Image Segmentation Model of described use and described initial profile line, each pixel is adjacent carry out curve EVOLUTIONARY COMPUTATION to described initial profile line, are specially:

When the iteration result of calculating when described curve evolvement reaches preset times for convergence or iterations, the curve evolvement result of calculation obtaining.

6. a PET cardiac image Myocardial Contour segmentation device, is characterized in that, comprising:

Initial profile line cutting unit, carries out coarse segmentation for the PET cardiac image to receiving, and obtains the initial profile line of myocardial region;

Fractional order setting unit, is set to fractional order for the derivative exponent number α in the length of curve regular terms of the function representation of level set Image Segmentation Model;

Curve evolvement computing unit, carries out curve EVOLUTIONARY COMPUTATION for using described level set Image Segmentation Model and choosing at least three pixels that on described initial profile line, each pixel is adjacent to described initial profile line;

Outline line acquiring unit, for the outline line using the curve evolvement result of calculation obtaining as described PET cardiac image Myocardial region.

7. device according to claim 6, is characterized in that,

Described level set Image Segmentation Model comprises unskirted active profile C-V(Chan-Vese) model or based on region extendible matching RSF(Region-Scalable Fitting) model.

8. device according to claim 6, is characterized in that,

Described derivative exponent number α is greater than 0 to be less than 2 and be not equal to 1 rational number, irrational number or plural number.

9. device according to claim 6, is characterized in that, described initial profile line cutting unit, is specially:

Use iteration threshold method to carry out coarse segmentation to the PET cardiac image receiving.

10. device according to claim 6, is characterized in that, described curve evolvement computing unit is specially:

When the iteration result of calculating when described curve evolvement reaches preset times for convergence or iterations, the curve evolvement result of calculation obtaining.

Images